
MR MILNE ON THE PARALLEL ROADS OF LOCHABER. 405 
the water sunk 14 feet, the discharge must have ceased at the east end; and that 
it henceforward would go on at the west end, probably near the mouth of Glen 
Glaster. At every other place, the rocky mountain sides rise so high, as to pre- 
clude the possibility of overflow or attrition. 
Keeping these principles in view, let us suppose that the detrital matter 
which blocked up the lower parts of Glen Roy extended a very little to the east 
of the mouth of Glen Glaster. How easy it is to suppose that this detritus was 
scooped away, so as to allow of the recession of the waters westward, and of their 
flowing round the east jaw of Glen Glaster, and on towards the head of that glen, 
from which they would descend to Glen Spean? For this purpose, it is not ne- 
cessary to suppose, that there was any lowering of the supposed barrier in level. 
even by asingle foot. All that is required is the scooping or wearing away of the 
detritus, so as to allow of the extension of the lake a little to the westward ;—a 
few yards would be sufficient. As the discharge at this first sinking, must have been 
at the west end, it is fair to infer that the wearing away of detritus took place 
there; and when once a flow of water was established through detrital matter. 
the process of removal would go on rapidly, so as to allow of repeated sinkings of 
the lake, till it reached the water shed at the head of Glen Glaster, the rocky 
nature of which would for a time stop any farther sinking, and thus allow of the 
formation of shelf 3. 
According to the foregoing views, we see how the waters would, by suc- 
cessive steps, sink from shelf 2 to shelf 3, and. after entering Glen Glaster, form 
a marking on both of its sides. We see, also, that the same removal of detritus 
which allowed the formation of shelf 3 in that glen, would allow also the exten- 
sion of it on Bohantine Hill, beyond the point where shelf 2 terminates. 
Whilst this process of attrition was going on in Glen Roy, there need have 
been no contemporaneous change in the blockage of Glen Collarig. But there 
also, at some time or other, a similar scooping out of detritus must have taken 
place, to allow of the extension of shelf 3 beyond the point where shelf 2 termi- 
nates. 
Nor is it difficult to conceive, how this removal of detritus was effected. Thus, 
in Glen Collarig, there are, on both sides of the glen, burns of considerable size 
and power (from the steepness of their channels) which flowed into the lake. 
There are three of them, which now descend in that part of the glen marked by 
shelves 2 and 3. If the detritus which formed the blockage in the lower part of 
the valley consisted of the same loose sand and gravel which now abounds there, 
forming cliffs from 70 to 80 feet high, nothing is more easy or natural than the 
scooping of it out, by such means. 
The same observations apply to the blockage in Glen Roy, which, to prevent 
the waters when at shelf 2 flowing into Glen Glaster, must have been near the 
mouth of Glen Collarig, called Gap in the maps, out of which, from the number of 
streams in it, a considerable current had flowed. 
